Biological Journal of the Linnean Society, 2009, 98, 278–290. With 3 figures Contrasting patterns of genetic structure in Caryocar (Caryocaraceae) congeners from flooded and upland Amazonian forestsbij_1287 278..290 Downloaded from https://academic.oup.com/biolinnean/article-abstract/98/2/278/2448039 by guest on 03 June 2020 ROSANE G. COLLEVATTI1,2*, LÉLIA C. T. LEOI1, SUEANE A. LEITE1 and ROGÉRIO GRIBEL3 1Pós-Graduação em Ciências Genômicas e Biotecnologia, Universidade Católica de Brasília. Brasília, DF 70790-160, Brazil 2Pós-graduação em Ecologia e Evolução, Departamento de Biologia Geral, Universidade Federal de Goiás, CP 131, Goiânia, GO 74001-970, Brazil 3Laboratório de Genética e Biologia Reprodutiva de Plantas, Coordenação de Pesquisas em Botânica, Instituto Nacional de Pesquisas na Amazônia/INPA, Manaus, AM 69099-000, Brazil Received 24 February 2009; accepted for publication 23 April 2009 In the present study, we compare the genetic structure of a flooded forest tree Caryocar microcarpum and a terra firme forest tree Caryocar villosum in the lower Rio Negro region and test the hypothesis that the Rio Negro, the largest tributary on the left bank of the Amazon River, has been acting as a geographical barrier to gene flow between populations from the left and right banks. Seventeen adult individuals on the left bank and 27 on the right bank of Rio Negro were sampled for C. microcarpum, whereas 27 on the left and 20 on the right bank were sampled for C. villosum. Two chloroplast DNA regions were sequenced: the intron of trnL gene and the intergenic region between psbA and trnH genes; and all individuals were genotyped using ten microsatellite loci. The trnL intron and psbA-trnH intergenic spacer generated fragments of 459 bp and 424 bp, respectively. For C. microcarpum, six haplotypes were identified for trnL and seven for psbA-trnH. By contrast, only one haplotype was found for C. villosum for both sequences. The results obtained showed that the Rio Negro has not been a barrier to gene flow by pollen and seeds for either species. No genetic differentiation and a high migration rate between populations from the left and right banks of the Rio Negro were detected for the chloroplast sequences and nuclear microsatellites, for both C. villosum and C. microcarpum. Although the two analysed sequences showed a sharp topology difference, both indicated that multiple lineages may have contributed to the origin of C. microcarpum populations in the Rio Negro basin. Nevertheless, for C. villosum, from terra firme, the results obtained may provide evidence of a recent expansion of one maternal lineage from an ancient relic population surviving in one of the few moist forest refuges of the Guiana Shield during extended droughts of the glacial periods. We hypothesize that the contrasting environments colonized by this congener pair may have played an important role in shaping the genetic structure of both species. © 2009 The Linnean Society of London, Biological Journal of the Linnean Society, 2009, 98, 278–290. ADDITIONAL KEYWORDS: Amazonia – Caryocar microcarpum – Caryocar villosum – Caryocaraceae – cpDNA – microsatellites – Neotropical tree – population genetic structure – riverine barrier hypothesis. INTRODUCTION episodes of the Tertiary and Pleistocene (Burnham & Graham, 1999; Vélez et al., 2006), with large-scale The distribution and composition of vegetation in the replacement of the tropical rain forest by drought- Amazon have been deeply influenced by the drier tolerant dry forests and savannas mainly in southern, central, and eastern Amazonia (Hooghiemstra & Van *Corresponding author. E-mail: [email protected] der Hammen, 1998; Pennington et al., 2000, 2004; 278 © 2009 The Linnean Society of London, Biological Journal of the Linnean Society, 2009, 98, 278–290 GENETIC STRUCTURE OF TWO AMAZONIAN CARYOCARACEAE SPECIES 279 Van der Hammen & Hooghiemstra, 2000; for an alter- analysis of the role of historical and contemporary native view, see Colinvaux, Oliveira & Bush, 2000). events on current species distributions and genetic The long series of ice ages may have led to extinctions structure. For plants, the analysis of nuclear and but, concomitantly, may have stimulated the evolu- chloroplast genomes may also clarify the relative tion and speciation of other groups (Van der Hammen, importance of pollen and seed flows on population 1974; Hooghiemstra & van der Hammen, 2004). The structure (McCauley, 1995; Schaal et al., 1998; Coll- altitudinal gradient may have allowed the invasion evatti, Grattapaglia & Hay, 2003). Moreover, analysis of montane plant species into the basin during the of different regions of the chloroplast genome, which cooler periods and the evolution and adaptation of display different mutation rates, may provide addi- other groups (Hoorn et al., 1995; Hooghiemstra & tional insights about the evolution and historical Downloaded from https://academic.oup.com/biolinnean/article-abstract/98/2/278/2448039 by guest on 03 June 2020 Van der Hammen, 1998, 2004; Colinvaux et al., 2000; spread of populations (Soltis & Soltis, 1998). The Urrego et al., 2006). Proxy data also indicate a genetic structure of the nuclear genome may be dynamic changing on the Amazon River and its tri- caused by historical and contemporary gene flow, by butaries, especially subsequent to the Miocene differential selection among habitats, genetic drift, (Hooghiemstra & Van der Hammen, 1998). On the and the mating systems that determine inbreeding other hand, sea level elevations during the Tertiary, effects on population differentiation (Wright, 1931). especially in the Miocene, caused the flooding of Nevertheless, genetic structure of the organelle extensive Amazonian lowlands, isolating populations genome, usually uniparentally inherited, is more at higher elevations, mainly in the Guianan Shield, affected by historical relationships and gene flow and Brazilian Shield, and the base of the eastern slope of by demographic fluctuations caused by historical the Andes, changing the landscape and affecting events such as glaciations and climatic fluctuations species distribution (Webb, 1995). over a geological time scale (Avise, 1994; Schaal et al., Analyses of paleodunes show that the Rio Negro 1998). Additionally, because of the haploid nature and may have attained lower levels during the glaciation mode of inheritance, the effective population size of periods of the Tertiary (Carneiro-Filho et al., 2002) the chloroplast genome is one-half the size of the and paleogeographic reconstruction has revealed nuclear genome, leading to a stronger effect of genetic Quaternary changes in the Rio Negro channel (Silva drift on population genetic structure (Ennos, 1994). et al., 2007). Hence, populations on both banks of the Caryocar villosum Aublet (Caryocaraceae), popu- Amazon River itself and its tributaries may have been larly known as piquiá, is a low-density widely- closer, or in contact, during some periods of the Qua- distributed Amazonian emergent tree species, up to ternary. Nevertheless, the riverine barrier hypothesis, 50 m tall, in the upland (terra firme) forests, whereas which had its own origin in Wallace’s paper ‘On the Caryocar microcarpum Ducke, known as piquiarana monkeys of the Amazon’ (Wallace, 1852) has been (false piquiá in the Tupi language), is a 25-m tall, raised to explain the megadiversity of Amazonia habitat-specific tree, growing in the seasonally- (Haffer, 1997). The main prediction of this hypothesis flooded blackwater forest (igapó forest), widespread is that the Amazon River and its major tributaries throughout the Guianas and Northern Amazonia have acted as geographical barriers, leading to vicari- (Prance & Freitas da Silva, 1973). The main goal of ance speciation in many groups. Another prediction of the present study was to compare the genetic struc- this hypothesis is the occurrence of higher genetic ture of this congener pair adapted for contrasting differentiation among populations on opposite river forest habitats in the lower Rio Negro region. Addi- banks. The relationship of dispersal and riverine bar- tionally, we tested the hypothesis that the Rio Negro riers is not so clear cut (Colwell, 2000). For many is a relevant geographical barrier for pollen and seed groups, riverine barriers could not account for the flows for both species. The results are discussed in the observed patterns of species richness, geographical light of the climatic and hydrographic changes that distribution, and differentiation (Ateles, Primata: occurred during the Plio-Pleistocene period in the Collins & Dubach, 2000; mammal species: Patton, Da lower Rio Negro region. The genetic structure of C. Silva & Malcolm, 2000; avian species: Bates, Haffer villosum and C. microcarpum was studied based on & Grismer, 2004; passerine birds: Hayes & Sewlal, the polymorphism at two chloroplast DNA regions 2004), but could account for other groups (frogs and on ten nuclear microsatellite loci. and mammals: Gascon, Malcolm & Patton, 2000; mammals: Patton et al., 2000; Riodinid butterflies: Hall & Harvey, 2002; passerine birds: Hayes & MATERIAL AND METHODS Sewlal, 2004). Comparative analysis of nuclear and organelle POPULATIONS, SAMPLING, AND DNA EXTRACTION genomes, with different modes of inheritance, muta- The Rio Negro, 1700 km long, is the largest left bank tion, and evolutionary rates, may provide a powerful tributary of the Amazon River and the largest black- © 2009 The Linnean Society of London, Biological Journal of the Linnean Society, 2009, 98, 278–290